Overexpression of the urokinase receptor mRNA splice variant uPAR-del4/5 affects tumor-associated processes of breast cancer cells in vitro and in vivo
- 180 Downloads
uPAR, the three-domain membrane receptor of the serine protease urokinase, plays a crucial role in tumor growth and metastasis. Several mRNA splice variants of this receptor have been reported. One of these, uPAR-del4/5, lacking exons 4 and 5, and thus encoding a uPAR form lacking domain DII, is specifically overexpressed in breast cancer and represents a statistically independent prognostic factor for distant metastasis-free survival in breast cancer patients. The aim of the present study was to examine the molecular and cellular properties of the encoded uPAR-del4/5 protein. To investigate the impact of the uPAR-del4/5 overexpression on in vitro and in vivo aspects of tumor progression (e.g., proliferation, adhesion, invasion, metastatic seeding, and/or metastatic growth), we combined the analysis of transfected cancer cell lines with a murine xenograft tumor model. Increased expression of uPAR-del4/5 in human cancer cells led to reduced adhesion to several extracellular matrix proteins and decreased invasion through MatrigelTM, while cell proliferation was not affected in vitro. Moreover, invasion of uPAR-del4/5 overexpressing cells was not altered by addition of urokinase, while that of uPAR-wild-type overexpressing cells was drastically increased. Accordingly, we observed that, in contrast to uPAR-wild-type, uPAR-del4/5 does not interact with urokinase. On the other hand, when overexpressed in human breast cancer cells, uPAR-del4/5 distinctly impaired metastatic dissemination and growth in vivo. We demonstrate that the uPAR-del4/5 mRNA splice variant mediates tumor-relevant biological processes in vitro and in vivo. Our results thus illustrate how tumor-specific alternative splicing can distinctly impact the biology of the tumor.
KeywordsBreast cancer Splice variant Adhesion Invasion Tumor xenograft uPAR
The serine protease urokinase-type plasminogen activator (uPA), its serpin-type inhibitor PAI-1, and its cell surface receptor uPAR (CD87), drastically influence tumor cell proliferation, migration, invasion, and metastasis [1, 2, 3]. Accordingly, these factors represent important biomarkers in a number of malignancies, including breast cancer [4, 5, 6].
uPAR is an ubiquitous 45–60 kDa glycoprotein, which is composed of three homologous domains (DI, DII, and DIII, from the amino- to the carboxy terminus). It is anchored to the plasma membrane by a glycosyl phosphatidyl inositol (GPI) moiety . uPAR exhibits features related to proteolysis as well as cell adhesion: on one hand, upon binding of its canonical ligand uPA, uPAR focuses the extracellular proteolytic network including plasmin and matrix metalloproteinases to the cell surface [2, 7, 8], on the other, it interacts with the adhesive extracellular matrix (ECM) protein vitronectin as well as with various integrins, thereby regulating cell adhesion and signaling [2, 8, 9]. For these reasons, uPAR is a major regulator of ECM remodeling and cell adhesion/migration [2, 8].
Recently, crystal structures of complexes between soluble uPAR (lacking the GPI anchor) and an antagonist peptide (interacting with the uPA binding site of uPAR) or the N-terminal fragment of uPA (ATF; harboring the uPAR-binding site) have been determined [10, 11]. These studies revealed that the three domains in uPAR are packed closely and form a cone-shaped central cavity which is wide open and displays a significant depth. The major contact areas between uPAR and these ligands are formed by DI and DII residues, respectively. Binding of uPA to uPAR dramatically increases the affinity of uPAR to vitronectin . The crystal structure of the ternary complex of soluble uPAR, ATF and the uPAR-binding domain of vitronectin, SMB, showed that there is no direct contact between SMB and ATF, whereas the uPAR-binding site of uPA occupies the central cavity of the receptor, SMB binds to the outer side of DI and the DI–DII linker region . Thus, uPA/uPAR-interaction may stabilize an active conformation of uPAR, leading to its high affinity binding to vitronectin.
Wild-type uPAR (uPAR-WT) is encoded by seven exons. Several mRNA splice variants of uPAR have been identified and their expression analyzed in human cells and tissues [14, 15, 16]. One of these splice variants, uPAR-del4/5, lacking exons 4 and 5, and thus encoding a uPAR form lacking domain DII, was found to be specifically overexpressed in breast and ovarian cancer cells . Besides, high uPAR-del4/5 mRNA expression levels were found to be significantly associated with short disease-free survival of breast cancer patients [15, 17, 18]. Moreover, uPAR-del4/5 mRNA is a highly sensitive, statistically independent prognostic factor for distant metastasis-free survival in untreated node-negative breast cancer patients .
In order to investigate the tumor biological effects of uPAR-del4/5 splice variant expression, we stably transfected MDA-MB-231 breast cancer cells with expression plasmids encoding either this variant or wild-type uPAR. By proliferation, adhesion, and invasion assays, we first analyzed the phenotype of the cell transfectants in vitro, and then monitored the impact of uPAR-del4/5 overexpression on experimental metastasis in a xenograft tumor model in mice.
Materials and methods
Cell culture and cell transfection
The human breast adenocarcinoma cell line MDA-MB-231 (American Type Culture Collection [ATCC], Manassas, VA) was cultured in DMEM (Gibco BRL, Eggenstein, Germany) supplemented with 10% (v/v) fetal calf serum (FCS), 10 mM HEPES, 0.55 mM l-arginine and 0.272 mM l-asparagine (Sigma-Aldrich, Saint-Louis, MO). In some experiments, MDA-MB-435, CAMA-1, MCF-7 breast cancer cells (ATCC), and OV-MZ-6 ovarian cancer cells , respectively, were also used. Cells were routinely checked to be free of mycoplasma.
Cloning of pRcRSV-derived expression plasmids encoding uPAR (-variants) has been reported elsewhere . Cells were transfected using Lipofectin® (Invitrogen, Karlsruhe, Germany). Cell transfectants were selected by addition of 1 mg/ml G418 to the cell culture medium (Gibco BRL). In an independent transfection experiment, clones of MDA-MB-231-uPAR-del4/5 cells were isolated by limited dilution.
Cells were washed with phosphate-buffered saline (PBS), pH 7.4, then lysed for 60 min at 4°C in Tris-buffered saline (TBS), pH 7.4, containing 1% (v/v) Triton X-100 and a protease inhibitor cocktail (“complete + EDTA”; Roche Diagnostics, Mannheim, Germany). Cell lysates were centrifuged at 13,000×g for 15 min at 4°C. Protein content was determined using the Micro BCATM protein assay reagent kit (Pierce, Stonehouse, UK). uPAR antigen was determined applying the IMUBIND uPAR ELISA kit #893 (American Diagnostica Inc., Stamford, CT). uPAR antigen levels in cell lysates are expressed as nanogram per milligram of total protein.
Proteins, separated by electrophoresis on 12% (w/v) polyacrylamide gels (SDS-PAGE), were transferred to polyvinylidene fluoride membranes (Millipore Corporation, Bedford, MA) in a semi-dry transfer device (Biometra, Göttingen, Germany). Membranes were incubated for 60 min in PBS, pH 7.4, containing 0.1% (v/v) Tween-20 (PBS-T) and 5% (w/v) dried skimmed milk, followed by an overnight incubation with the primary monoclonal antibodies (mAb) IIIF10 or IID7, both directed to uPAR, diluted in PBS-T supplemented with 1% (w/v) dried skimmed milk. The epitope-mapped mAb IIIF10 is directed against DI, mAb IID7 against DII of uPAR . After washes in PBS-T, binding of the antibodies was visualized by incubation of the membranes with a horseradish peroxidase-conjugated secondary goat Ab against mouse Ig (Jackson ImmunoResearch Lab, West Grove, PA), followed by chemoluminescence reaction using ECL (Amersham Biosciences, Little Chalfont, UK).
Cells were seeded in human fibronectin-coated (10 μg/ml, Becton-Dickinson, Heidelberg, Germany) 8-chamber glass slides (Permanox-type Lab-Tek slides; Nunc, Roskilde, Denmark) and cultured overnight. Cells were fixed in 4% (w/v) paraformaldehyde in PBS, pH 7.4, for 30 min at room temperature (RT). After several washes in PBS, cells were incubated for 1 h at RT in PBS containing 2% (w/v) bovine serum albumin (BSA), and then probed with mAb IIIF10 diluted in PBS, 1% (w/v) BSA, for 1 h at RT. Cells were washed and incubated for 45 min at RT in the dark with the secondary Alexa488-labeled goat-anti-mouse IgG (Sigma-Aldrich) diluted in the same buffer. After final washes with PBS, cells were mounted in PBS and fluorescence signal intensity evaluated by confocal laser scanning microscopy (CLSM).
Solid-phase uPA binding assay
The solid-phase uPA-ligand-binding assay was performed in a similar manner as described previously . Briefly, wells of microtiter plates (Maxisorb; Nunc) were coated overnight at 4°C with 100 μl of rec-uPAR1–277 (1 μg/ml). After blocking, 100 μl of sample containing the aminoterminal fragment (ATF) of uPA (250 ng/ml) and 50 μl of cell culture supernatants were added to the wells for 1 h at RT. Cell culture supernatants were derived from vector-control Chinese hamster ovary (CHO) cells or from transfected CHO cells overexpressing soluble forms of uPAR-WT or uPAR-del4/5 . After four washes, plates were incubated with biotinylated murine mAb #377 directed to the kringle domain of human uPA (American Diagnostica Inc.) for 1 h at RT. After washes in PBS, avidin-coupled peroxidase was added for 1 h at RT. Finally, wells were washed and binding of avidin-peroxidase detected by addition of 3,3′,5,5′-tetramethlybenzidine (TMB; 1 mg/ml) and 0.003% (v/v) H2O2 in 0.1 M sodium acetate buffer, pH 6.0. After termination of the reaction with 0.5 M H2SO4, absorbance was measured at 450 nm.
Cell-based in vitro assays
Cell proliferation assays
Transfected cells were seeded in a 24-well plate at a density of 15,000 cells per well, and incubated for 48–96 h at 37°C. After washes in PBS, cells were detached using PBS, 0.05% (w/v) EDTA, and living cells counted in a hemocytometer upon trypan blue exclusion under a light microscope.
Cell adhesion assays
96-well plates were coated overnight at 4°C with vitronectin (1 μg/well), fibronectin (1 μg/well), and collagen type I or IV (1 μg/well), all diluted in PBS. Cells were resuspended in culture medium containing 0.5% (w/v) BSA, seeded at a density of 40,000 cells/well, and allowed to adhere for 2 h at 37°C. Non-adherent cells were removed by washing with PBS and number of adherent cells quantified by a hexosaminidase activity assay. For this, cells were incubated with p-nitrophenyl-N-acetyl-β-d-glucosaminide (Sigma-Aldrich) diluted to 15 mM in a 100 mM sodium citrate buffer, pH 5.0, 0.5% (v/v) Triton X-100, for 90 min at 37°C. The reaction was terminated by the addition of stop buffer (0.2 M NaOH, 5 mM EDTA) and the optical density recorded at 405 nm.
Cell invasion assays
Invasion assays were performed using Transwell inserts (8 μm pore size; Becton-Dickinson). Basement membrane complex growth factor reduced MatrigelTM (11.3 mg/ml) was diluted 1:24 in cold PBS and applied to the upper side of the insert. After drying for 24 h in a laminar hood, inserts were rehydrated with FCS-free DMEM, 0.1% (w/v) BSA. Cells were resuspended in culture medium and seeded in the upper chamber of the device at a density of 50,000 cells/chamber. In some experiments, 2 μg of proteolytically active high-molecular-weight uPA (ProSpec-Tany TechnoGene, Rehovot, Israel) were added to the upper chamber. The lower chambers were filled with 750 μl DMEM supplemented with 10% (v/v) FCS as a chemoattractant. After 24 to 48 h of incubation at 37°C, MatrigelTM and non-invaded cells, located on the upper side of the insert, were removed with tissue paper, whereas invaded cells on the lower side of the insert were fixed, stained using Diff-Quick (Dade Behring AG, Switzerland), and counted under a light microscope.
Experimental animal model
MDA-MB-231 cells are tumorigenic and invasive cells, which upon intravenous injection in mice colonize to the lungs . Transfected MDA-MB-231 cells were genetically tagged with the lacZ gene, allowing X-Gal staining, as reported previously . Pathogen-free female athymic (nu/nu) mice (4–6 weeks old) were obtained from Charles River Laboratories (Sulzfeld, Germany). Mice were allocated to four groups and intravenously inoculated (tail vein) with 1 × 106 (a) control MDA-MB-231-vector cells (n = 11), (b) batch-transfected MDA-MB-231-uPAR-del4/5 cells (n = 10), (c) cloned MDA-MB-231-uPAR-del4/5 cells (n = 10), or (d) batch-transfected MDA-MB-231-uPAR-WT cells (n = 9). At day 35 post-injection, animals were killed, lungs isolated and the lacZ-tagged tumor cells stained with X-Gal (Roche Diagnostics) . Thereafter, blue-stained metastatic foci were counted and their size measured, considering nodules with diameter >0.18 mm as macrometastases.
Data are expressed as mean ± SE (standard error) of the indicated number of experiments. Group differences and P values were calculated by employing the Mann–Whitney test. A P value <0.05 was considered statistically significant.
Generation of uPAR overexpressing breast cancer cell lines
In vitro phenotype of uPAR-del4/5 overexpressing breast cancer cells
uPAR-del4/5 overexpression reduces the adhesive capacity of human breast and ovarian cancer cells in vitro
(% of the number of adherent vector-transfected cells)
86.5 ± 5.5*
80.0 ± 23.1*
83.7 ± 11.7*
89.9 ± 17.7
87.6 ± 17.3*
74.5 ± 16.7*
66.7 ± 25.4*
82.5 ± 36.3
71.7 ± 17.9*
57.9 ± 28.3*
29.8 ± 14.8*
39.4 ± 21.8*
93.9 ± 4.0*
92.8 ± 5.3*
92.5 ± 2.8*
98.7 ± 6.6
89.7 ± 6.4*
85.8 ± 11.2*
92.5 ± 6.9*
95.9 ± 13.7
In vivo phenotype of uPAR-del4/5 overexpressing breast cancer cells
Aberrant mRNA splicing is a common feature of malignant disorders. Whether this process results from a general dysregulation of tumor cell functions, and thus represents a by-product of cellular transformation, or whether increased aberrant splicing contributes to the malignant phenotype of cancer cells, is a not yet conclusively answered question [30, 31]. Still, a growing number of evidence demonstrates that tumor-associated mRNA splicing may lead to cancer cell-specific production of new proteins displaying unique functions, which are crucially affecting cancer development. Mis-spliced mRNAs, which escape the so-called nonsense-mediated mRNA degradation, normally lead to synthesis of truncated, mutated, misfolded, and/or unstable proteins. However, by alternative splicing, it is also possible that certain protein domains are deleted from the encoded protein, which may result in loss or gain of function . In a recent study on splice variants of the human Adhesion family of G protein-coupled receptors, more than half of the identified splice variants appeared to code for functional proteins, lacking one or more extracellular domains, which take part in protein–protein interactions, without affecting the seven transmembrane region. Thus, alternative splicing apparently can influence interaction of these receptors with other proteins .
The recently identified, tumor-associated variant of uPAR, uPAR-del4/5, displays the specific deletion of the complete DII of three-domain uPAR . In fact, uPAR-del4/5 protein, consisting of domains DI + DIII fused together, exhibits unique characteristics significantly affecting the behavior of malignant cells: upon analyses of the impact of uPAR-del4/5 overexpression in human MDA-MB-231 breast cancer cells, we observed a significantly impaired cell adhesion and invasion in vitro, along with a distinctly decreased colonization of lungs (in terms of metastatic nodule occurrence, number, and size) in an experimental metastasis mouse model system.
Binding of uPA to its cell surface receptor uPAR involves all three domains of uPAR [2, 10]. As demonstrated by a sensitive solid-phase ligand-binding assay (Fig. 4) and by an uPA/uPAR-complex ELISA (data not shown), uPAR-del4/5 does not interact with uPA. Consequently, we observed that addition of uPA to uPAR-del4/5 overexpressing cells—in contrast to uPAR-WT overexpressing cells—had no effect on cell invasion through MatrigelTM (Fig. 3).
uPA/uPAR-interaction does not only focus proteolytic activity to the tumor cell surface, but is also a prerequisite for the binding of another uPAR ligand, the ECM protein vitronectin . Furthermore, interaction of uPA with uPAR is a major determinant of physical and functional interactions of uPAR with a number of integrins [2, 9] and is also involved in other tumor-associated functions, e.g. regarding tumor cell survival and angiogenesis [2, 33]. Obviously, uPAR-del4/5 is not capable of modulating these uPA-dependent, tumor-associated functions of uPAR-WT by competing for uPA binding. Still, overexpression of uPAR-del4/5 resulted in a moderate, significantly decreased cell adhesion to an array of adhesive ECM proteins (Fig. 2, Table 1). uPAR-del4/5 also strongly affected cellular invasion in vitro (Fig. 3), a process that depends on both extracellular proteolysis and cell adhesion/detachment capacities. In addition, overexpression of the uPAR-del4/5 splice variant led to a major reduction of experimental metastasis in vivo, compared to both vector control and uPAR-WT overexpressing cells (Fig. 5). It is tempting to speculate that such a phenotype might reflect the capacity of uPAR-del4/5 to behave as a dominant-negative receptor, e.g. via competition with endogenous uPAR-WT for binding to certain integrins such as α5ß1 or αvß3 and/or via an additional uPAR-WT independent process.
uPAR-del4/5 mRNA is a statistically independent prognostic factor for distant metastasis-free survival in breast cancer patients. High levels of uPAR-del4/5 mRNA are associated with a poor prognosis of patients [15, 17, 18]. However, the results of the present article, especially regarding the reduced invasive capacity of uPAR-del4/5 overexpressing cells in vitro (Fig. 3) and the reduction in lung colonization of these cells in the in vivo experimental metastasis animal model (Fig. 5) rather advert to the conclusion that the uPAR-del4/5 protein acts as a forceful tumor suppressor. On the one hand, high expression of uPAR-del4/5 mRNA might reflect the malignant status of the cancer disease, rather than the functions arising thereof at the protein level. On the other hand, as documented for the adverse effects of different PAI-1 concentrations on tumor angiogenesis , uPAR-del4/5 protein, depending on the height of its expression level, may display both agonistic and antagonistic properties. Therefore, further extensive studies using clones expressing uPAR-del4/5 protein to a different extent (low vs. high expression) and/or variations of uPAR-del4/5: uPAR-WT protein ratios may help to explore this latter possibility.
The metastatic process requires a complex sequence of events . Invasive tumor cells, originating from an in situ cancer surrounded by an intact basement membrane, degrade ECM proteins, induce reversible changes in cell–cell and cell–ECM adherence, and migrate through the ECM. The metastasizing cells then enter circulation, extravasate and disseminate to eventually colonize at a distant site, finally leading to micrometastases and angiogenic metastases. The experimental animal metastasis model used in the present study focuses on the later events of the metastatic process, i.e. metastatic seeding and growth. Here, uPAR-del4/5 overexpression did not only reduce occurrence and number of metastatic nodules, but also the percentage of macrometastases (Fig. 5), which indeed indicates an impact of the uPAR splice variant on extravasation of circulating tumor cells and/or tissue colonization, as well as on metastatic tumor growth.
In conclusion, although several uPAR mRNA splice variants have been described, some being clearly associated to disease [14, 15, 16], no experimental analysis of their cellular properties was reported so far. The present study, combining in vitro and in vivo experimental approaches, clearly demonstrates tumor biologically relevant effects mediated by uPAR-del4/5 in human breast cancer and also ovarian cancer cells. Thus, production of uPAR-del4/5 might represent a striking example for tumor-associated alternative splicing leading to modified protein species strongly modulating tumor biological processes.
This study was supported in part by grants provided by the Deutsche Krebshilfe e.V., Germany (Grant No. 106 162) to MK and VM, and by the Framework Programme 7 project HEALTH-2007-201279, Microenvimet to AK. We are grateful to Sabine Creutzburg, Katja Richter, and Antje Zobjack for excellent technical assistance.
Conflicts of interest statement
- 9.Reuning U, Magdolen V, Hapke S, Schmitt M (2003) Molecular and functional interdependence of the urokinase-type plasminogen activator system with integrins. Biol Chem 384:1119–1131Google Scholar
- 15.Luther T, Kotzsch M, Meye A, Langerholc T, Fussel S, Olbricht N, Albrecht S, Ockert D, Muehlenweg B, Friedrich K, Grosser M, Schmitt M, Baretton G, Magdolen V (2003) Identification of a novel urokinase receptor splice variant and its prognostic relevance in breast cancer. Thromb Haemost 89:705–717PubMedGoogle Scholar
- 18.Kotzsch M, Sieuwerts AM, Grosser M, Meye A, Fuessel S, Meijer-van Gelder ME, Smid M, Schmitt M, Baretton G, Luther T, Magdolen V, Foekens JA (2008) Urokinase receptor splice variant uPAR-del4/5-associated gene expression in breast cancer: identification of rab31 as an independent prognostic factor. Breast Cancer Res Treat 111:229–240PubMedCrossRefGoogle Scholar
- 21.Goretzki L, Bognacki J, Koppitz M, Rettenberger P, Magdolen V, Creutzburg S, Hammelburger J, Weidle UH, Wilhelm O, Kessler H, Graeff H, Schmitt M (1997) Quantitative assessment of interaction of urokinase-type plasminogen activator and its receptor (CD87) by use of a solid-phase uPA-ligand binding assay. Fibrinol Proteol 11:11–19CrossRefGoogle Scholar
- 25.Beaufort N, Leduc D, Eguchi H, Mengele K, Hellmann D, Masegi T, Kamimura T, Yasuoka S, Fend F, Chignard M, Pidard D (2007) The human airway trypsin-like protease modulates the urokinase receptor (uPAR, CD87) structure and functions. Am J Physiol Lung Cell Mol Physiol 292:1263–1272CrossRefGoogle Scholar
- 28.Krüger A, Soeltl R, Lutz V, Wilhelm OG, Magdolen V, Rojo EE, Hantzopoulos PA, Graeff H, Gänsbacher B, Schmitt M (2000) Reduction of breast carcinoma tumor growth and lung colonization by overexpression of the soluble urokinase-type plasminogen activator receptor (CD87). Cancer Gene Ther 7:292–299PubMedCrossRefGoogle Scholar
- 29.Setyono-Han B, Sturzebecher J, Schmalix WA, Muehlenweg B, Sieuwerts AM, Timmermans M, Magdolen V, Schmitt M, Klijn JG, Foekens JA (2005) Suppression of rat breast cancer metastasis and reduction of primary tumour growth by the small synthetic urokinase inhibitor WX-UK1. Thromb Haemost 93:779–786PubMedGoogle Scholar